1. Field of the Invention
The invention relates to accessing segments of a media object over a network. More particularly, the invention provides a system and method for determining a preliminary transmission period for a segment of a media object.
2. Description of the Related Art
Streaming media is a method of making audio, video, and other multimedia data available in real-time, over a network. FIG. 1 illustrates a typical streaming media system. If the source of the streaming media is a live performance, the performance is often recorded as a data signal by an input device, such as, for example, a microphone 102 or a camera 104. The input device then sends the data signal to an encoding system 106 which converts the data signal into a digital form and compresses the digital signal into a streamable data object 107. The streamable data object 107 may be sent to a content creation station 108 for editing or may alternatively be sent directly to a streaming media server 110. The content creation station 108 includes content creation software, such as video editing software, that allows the user to modify the streamable data object 107. The content creation station 108 can also be used to independently create a streamable data object 107. After being modified or created by the content creation station 108, the streamable data object 107 may then be transmitted to the streaming media server 110.
Once the streamable data object 107 is located on the streaming media server 110, the streamable data object 107 is made available for further distribution over a network 100, such as the Internet, to one or more client computers 112. The client computer 112 generally requests access to the streamable data object 107, begins receiving portions of it, and begins rendering the streamable data object 107 into a multimedia presentation for an end user.
One problem, however, related to transmitting streamable data objects over a network is that the client computer 112 must determine when it has sufficient data to begin the presentation or playing of the streamable data object 107. One solution has been to require the client computer 112 to store the whole presentation in memory before beginning the display of the presentation. However, this approach consumes significant system resources and creates unacceptable delay between the user's initial request for the media and the start of the presentation. Also, the client computer 112 may not have the system capacity to store the whole streamable data object 107.
Under another approach implemented by some existing systems, the client computer 112 starts playing the streamable data object 107 as soon as it is received. Of the problems with this approach, perhaps the most unacceptable is the periodic freezing or delay during playback caused by a lack of sufficient data to render the streamable data object 107. Bandwidth intensive segments of the streamable data object 107 are often responsible for such. To ensure a seamless presentation, the client computer 112 may require more data from the streaming media server 110 than has been transmitted over the network 100. But the amount of data a client computer 112 can receive in any period of time is restricted by the type of network communication it uses.
For example, assuming that the network 100 is the Internet, it is common for a client computer 112 to use a modem to connect to the Internet. However, modems have limited bandwidth capabilities compared to other transmission mediums, i.e., television, radio, Ethernet. The bandwidth of a communication device is defined by the amount of data that can be communicated over a specified time. A 28.8 Kbps (kilobits per second) modem can transmit or receive around 29,000 bits per second.
Depending on the content of a streamable data object 107, the rate at which data is consumed during presentation of the streamable data object 107 may be much higher than the transmission rate for a particular time interval. In such cases, the client computer 112 of a conventional streaming media system simply waits in the middle of the presentation of the streamable data object 107 for the transmission of additional data, creating an unpleasant interruption for the user.
FIG. 2 is a high level block diagram representing the data transmission process of the prior art. Starting at a state 200, the client computer 112 (FIG. 1) requests the streamable media server 110 (FIG. 1) to start transmitting data of the streamable data object 107. Next, at a state 202, the streaming media server 110 starts sending data of the streamable data object 107 (FIG. 1) to the client computer 112. Moving to a state 204, the client computer 112 starts displaying the presentation of the transmitted data. Continuing to a state 206, the client computer 112 stops the presentation in order to wait for further data. Lastly, at a state 208, the client computer 112 finishes displaying the presentation. Although only one interruption is indicated in the flowchart of FIG. 2, the client computer 112 may suffer from multiple interruptions due to having insufficient data at various times in the rendering process.
For further example, referring to FIGS. 3 and 4, a consumption graph for an exemplary rendering of a streamable data object 107 is described. FIG. 4 shows the target bit rate of the client computer 112 in addition to the consumption graph of FIG. 3. The consumption graph illustrates that the number of bits that are consumed during presentation of a streamable data object 107 varies over time. During the first one second interval of the presentation, the client computer 112 (FIG. 1) consumes 30 K bits of data. However, during the next second, the client computer 112 consumes only 10 K bits of information. Table 1, based on the consumption graph in FIGS. 3 and 4, further illustrates a situation wherein the total number of bits that are needed by the client computer 112 to maintain an uninterrupted playback is larger than the total number of bits it has received at multiple points in time.
TABLE 1TimeBitsTotal BitsTotal Bits(seconds)Bits NeededTransferredNeededTransferred130K20k30K20K210K20K40K40K330K20K70K60K410K20K80K80K520K20K100K 100K 610K10K110K 110K 
As can be seen from Table 1 and FIG. 2, after one second, in order to seamlessly present the streamable data object 107, the client computer 112 needs 30 K bits; however, the client computer 112 has only received 20 K bits of the streamable data object 107. Therefore, the client computer 112 has to wait for further data before starting the presentation. Further, after the presentation is started, the client computer 112 has to halt again because the presentation requires 70 K bits of data at three seconds, but the client computer has only received 60 K bits. This halt manifests itself, in existing systems, by freezing the presentation.
To attempt to overcome this difficulty, some systems transmit a portion of a streamable data object before playback starts. However, all multimedia presentations are not always played from the beginning, and transmitting an initial portion of a streamable data object does not help to address such situations.
The client computer 112 typically has a rendering program (not shown) that allows the user to “seek” to the middle or any number of secondary positions in the presentation before or during playback. After starting to play at the secondary position, the client computer 112 may require more data for the presentation than has been transmitted. Thus, the client computer 112 must pause and wait for the transmission of additional data. One solution to this problem has been to estimate the initial transmission buffer for the secondary position as being equal to or slightly greater than the initial transmission buffer for the whole streamable data object. However, this approach is merely a rough estimate and cannot guarantee that the client computer 112 will render the presentation without further interruptions and additional waits for further data. Thus, this approach does little to ensure a smooth, uninterrupted playback of a media object.
Variations in rendering requirements (data consumption over time) are more apparent in low bandwidth media systems, such as MacroMedia Shockwave Flash 2.0. MacroMedia Shockwave Flash 2.0 is one type of content creation system that generates television-like cartoons for business, entertainment or education. In these low bandwidth media systems, each of the graphical objects (“characters”) of the presentation are well defined, and multimedia presentations are often developed wherein these characters move from segment to segment or from frame to frame. Accordingly, the rendering requirements of the presentation vary greatly, generally becoming quite high when a new character is introduced. Due to this variation, the amount of data that would ensure uninterrupted playback if received prior to rendering the streamable data object 107 from its beginning, may differ widely from the amount needed to ensure uninterrupted playback from a secondary position, such as the middle of the streamable data object 107.
Therefore, there is a need for a system and a method that provides for the uninterrupted presentation of streamable data objects, including those streamable media objects imposing rendering requirements that vary over time. Moreover, the system and the method should not require the client computer 112 to store the entire streamable date object before starting the presentation. Further, the system and the method should provide for the uninterrupted rendering of a streamable data object 107 for each of the possible playback positions of the streamable data object 107.